A Yield Line Component Method For Bolted Flange Connections-1

ENGINEERING JOURNAL / SECOND QUARTER / 2011 / 93
A Yield Line Component Method

for Bolted Flange Connections
BO DOWSWELL
ABSTRACT
Bolted connections are often used in steel structures to transfer of tension loads into wide flange members. The strength of these connections
is determined with a prying action design procedure (outlined in the 13th edition AISC Steel Construction Manual) that checks the limit states
of bolt tension rupture and bending of the flange. This procedure is valid only for fittings with limited bolt spacing and limited edge distance.
This paper discusses a method to determine the local flange bending strength of a wide flange member using the yield line method. The
proposed design method includes the effect of prying action on the bolts, and can be applied to many different connection configurations,
including connections with large bolt spacing and edge distances and connections with web stiffeners. Comparisons with test data from 10
independent research projects will be used to verify the accuracy of the proposed method.
Keywords: bolted tension connections, hangers, prying action.
M
any bolted connections in steel structures rely on the
transfer of tension loads into wide flange members
as shown in Figure 1. The strength of these connections is
determined with the prying action design procedure in the
Steel Construction Manual (AISC, 2005a), hereafter re-
ferred to as the Manual, which checks the limit states of bolt
tension rupture and bending of the flange. The procedure in
the Manual is valid only for fittings with limited bolt spac-
ing and edge distance such as clip angles at the end of a
beam.
The Manual does not provide guidance on how to de-
termine the equivalent length of fittings with large edge
distances and bolt spacings. In practice, conservative as-
sumptions are often made. It is commonly assumed that the
tributary length per bolt is twice the distance from the center
of the bolt to the face of the supporting web. This method
is slightly conservative for calculating the elastic stress for
wide cantilever beams loaded at the free end (Young, 1989);
however, it is extremely conservative for calculating the
strength of flanges in bending.
In other cases, unconservative assumptions are some-
times made, where web stiffeners are provided to prevent
flange bending, and the stiffened flange is assumed ade-
quate to carry the applied loads with no further calculations.
However, tests have shown that flange bending is a common
failure mode for connections with web stiffeners (Packer
and Morris, 1977; Garrett, 1977; Ghassemieh et al., 1983;
Moore and Sims, 1986; Zoetemeijer, 1981).
This paper will discuss a method to determine the local
flange bending strength of a wide flange member using the
yield line method. The proposed design method includes the
effect of prying action on the bolts and can be applied to
many different connection configurations, including con-
nections with large spacings and edge distances and connec-
tions with web stiffeners. Comparisons with test data from
10 independent research projects will be used to verify the
accuracy of the proposed method.
Bo Dowswell, P.E., Ph.D., Principal, SDS Resources, Birmingham, AL. E-mail:
bo@sdsresources.com
Fig. 1. Bolted hanger connection.
093-116_ej2q_2011_2010-03r.indd 93 7/21/11 3:42 PM
94 / ENGINEERING JOURNAL / SECOND QUARTER / 2011
BACKGROUND
Prying Action
When bolts are loaded in tension, deformation of the con-
nected parts will cause an increase in bolt tension. This ad-
ditional bolt tension is the prying force, q, shown in Figure 2.
Designing for prying action involves checking the limit
states of bending of the fitting and tension rupture of the
bolts. The two limit states are interdependentfor a given
load, an increase in flange thickness leads to a lower prying
force on the bolt.
The moment diagram of half of the flange is shown in
Figure 2. The moment at the face of the web is always
equal to the plastic capacity of the fitting, but the moment
at the bolt line can be reduced if required to limit the pry-
ing force on the bolt. This behavior is accounted for in the
design method in the Manual. The background for the de-
sign method is provided by Astaneh (1985), Thornton (1985)
and Kulak et al. (1987). To calculate the available tensile
strength when the connection geometry is known, Equation
1 is applicable:

T B
t
t
B
a
c
j
(
,
\
,
(
+
( )
2
1
(1)
where
LRFD ASD
t
Bb
pF
c
u
4 44 .
t
Bb
pF
c
u
6 66 .

+ ( )
j
(
,
\
,
(

,
,
,
]
]
]
]

1
1
2
1 1 00
t
t
c
.
=

1
d
p
b
a
b b
d
b
2
j
(
,
\
,
(

j
(
,
\
,
(
a a
d
b
d
b b
= +
2
1.25 +
2
B = available tensile strength per bolt, kips
a = distance from the bolt centerline to the edge of the
fitting, in.
b = distance from bolt centerline to the face of the web,
in.
d
b
= bolt diameter, in.
d = width of the hole along the length of the fitting, in.
F
u
= specified minimum tensile strength of connecting
element, ksi
p = tributary length of fitting per bolt, in.
t = thickness of the fitting, in.
The Yield Line Method
The yield line method was developed by Hognestad (1953)
and Johansen (1962) to determine the ultimate strength of
concrete slabs. It is an upper-bound solution based on the
principle of virtual work. One form of the upper-bound the-
orem of limit analysis states that a load calculated based on
an assumed mechanism will be greater than or equal to the
true limit load.
The yield line method requires the failure pattern to be
known prior to calculation of the collapse load. Many pat-
terns may be valid for a particular joint configuration. Be-
cause the collapse load is upper bound, the pattern that gives
the lowest load will provide results closest to the true failure
load. Therefore, selection of the proper yield line pattern is
important because an incorrect failure pattern will produce
unsafe results.
The collapse load is calculated assuming that a plastic
mechanism forms along each line of the chosen failure pat-
tern. To maintain equilibrium, the external work done by
the load moving through the virtual displacement, , must
equal the strain energy due to the plastic moment rotating
through virtual rotations,
i
. The virtual rotations are as-
sumed small, so
i
tan(
i
) sin(
i
). The influence of strain
hardening and membrane effects are not accounted for in
yield line analysis; therefore, there is potentially a large re-
serve capacity beyond the calculated collapse load.
Some yield line patterns will produce an equation for the
load in terms of known geometry, but most cases will require
any unknown dimensions to be determined by minimizing
the load with respect to the unknown dimension. To do this,
the load is differentiated with respect to the unknown di-
mension and set equal to zero. From this, an equation for the
unknown dimension can be determined and substituted into
the equation for the load.
The general procedure for deriving an equation based on
yield line analysis is as follows:
Fig. 2. Model for prying action design method.
093-116_ej2q_2011_2010-03r.indd 94 7/21/11 3:42 PM
Select a valid yield line pattern.
Determine the equation that describes the external
work done by the load moving through the virtual
displacement.
W
E
= P (2)
where
P = applied load
= virtual displacement
Determine the equation that describes the internal work
done by the rotations along the yield lines,
W M
I pi i
(3)
where
M
pi
= plastic moment capacity of yield line i
= m
p
L
i

i
= virtual rotation of yield line i
m
p
= plastic moment capacity per unit length of the
fitting
= F
y
t
2
/4
L
i
= length of yield line i
Set the external work equal to the internal work and
solve for the load. If required, minimize the load with
respect to unknown dimensions.
Traditionally, the prying action equations have been
derived using equilibrium methods (Kulak et al., 1987), but
the equations can also be derived using energy methods. To
show the similarity between the design method for prying
action and the yield line equations for flange bending, the
Manual equation for the required fitting thickness will be
derived for the case of an infinitely strong bolt. This exercise
will also show the validity of the yield line method for this
simple case.
Considering only one side of the connection in Figure 3,
the external work is
W
E
= T (4)
The internal work is
W
I
= m
p
(L
1
+ L
2
) (5)
where
L
1
= length of yield line 1
= tributary length per bolt, p
L
2
= length of yield line 2
= net tributary length per bolt, pd
Substitute L
1
= p and L
2
= p d into Equation 5 to get
W
I
= m
p
(2p d) (6)
For small angles, = /b'
W
b
m p d
I p

( )
2 (7)
Substitute m
p
= F
y
t
2
/4
W
F t
b
p d
I
y

( )
2
4
2 (8)
Set internal work equal to external work and solve for T
n
T
F t
b
p d
n
y

( )
2
4
2 (9)
The available LRFD strength is

T
F t
b
p d
n
y

( )
2
4
2 (10)
Rearrange Equation 10 and solve for the thickness of the
fitting
t
Tb
F p d
min
y

( )
4
2
(11)
Substitute = 0.90 into Equation 11
t
Tb
F p d
min
y

( )
4 44
2
.
(12)
Fig. 3. Yield line model for prying equation.
093-116_ej2q_2011_2010-03r.indd 95 7/21/11 3:42 PM
Dranger (1977)
The yield line pattern in Figure 5 was solved by Dranger
(1977), who determined the strength as a function of the un-
known dimension x:
P F t
x
b
c
x
n y
+
j
(
,
\
,
(
2
(15)
Dimension x was then determined by minimizing the load:
x bc (16)
If x from Equation 16 is substituted into Equation 15, the
nominal strength is

P F t
c
b
n y
2
2
(17)
where
c = a + b
Mann and Morris (1979)
Mann and Morris (1979) presented a yield line pattern
with circular corners as shown in Figure 6. The nominal
strength is
P F t
a p d
b
n y
+
+
j
(
,
\
,
(
2
2
(18)
Equation 18 defines the total connection strength, which
is twice the strength of each independent yield line pattern
forming on both sides of the column web. Mann and Morris
also suggested an equation similar to Drangers (1977) for
stiffened connections; however, no guidance was given on
how close the stiffener has to be to the bolt for that equation
to apply.
The LRFD version of the prying equation (on page 9-11 of
the Manual) is
t
Tb
pF
u
min
.

+
( )
4 44
1
(13)
Although the Manual procedure uses the ultimate tensile
strength, F
u
, for prying calculations, which was first sug-
gested by Douty and McGuire (1965) and more recently by
Thornton (1992), yield line analysis has traditionally utilized
the yield strength, F
y
. For comparison with the yield line
derivation, F
y
will be used here. Replacing F
u
with F
y
in
Equation 13, substituting = 1.0 for infinitely strong bolts
and substituting = 1 d/p, Equation 12 is obtained.
EXISTING SOLUTIONS
Yield line theory has been presented as a design method for
bolted connections in several publications, and many dif-
ferent yield line patterns have been proposed. A component
method, similar to the design method proposed in this pa-
per, is currently used in Europe (SCI, 1995; CEN, 2005) to
determine the column flange bending strength and the plate
bending strength in moment end plate connections.
Zoetemeijer (1974)
The equivalent length concept was first discussed by
Zoetemeijer (1974), who used a simplified solution to the
yield line pattern in Figure 4 to get an equivalent tributary
length per bolt of
p b
a p
e
+ + 2
5
8 2
(14)
where
p = spacing between bolts
Fig. 4. Yield line pattern from Zoetemeijer (1974). Fig. 5. Yield line pattern from Dranger (1977).
093-116_ej2q_2011_2010-03r.indd 96 7/21/11 3:42 PM
these tests, four connection types were identified, based
on the edge distance and stiffener configuration. These are
presented in Table 1. The specimen details are shown in
Table A1 of Appendix A.
The generalized experimental load-deflection curve is bi-
linear with a nonlinear transition point as shown in Figure 8.
There are four points of interest on the curve:
1. The proportional limit, where the curve transitions
from linear to nonlinear. The load at this deformation
may be of interest as a serviceability limit for connec-
tions that can allow only very small deformations. The
deformation at this point is
p
,

and the load is P
p
.
2. The point where the curve transitions from nonlinear
to linear at the second linear part of the curve. Loads
increased beyond this point are accompanied by large
deformations. The deformation at this point is
s
, and
the load is P
s
.
3. The point of 4-in. deformation. This is proposed here
as the serviceability limit. The deformation at this
point,
4
is 4 in., and the load is P
4
.
Zoetemeijer (1981)
Zoetemeijer (1981) presented a circular yield line pattern as
shown in Figure 7, which he described as a punching fail-
ure. For this pull-through mechanism, the prying force is
theoretically zero. The yield line solution predicts a nominal
strength of
P
n
= F
y
t
2
(19)
Thornton and Kane (1999) and Muir and Thornton
(2006)
Thornton and Kane (1999) and Muir and Thornton (2006)
published the following equation, which provides the aver-
age equivalent length per bolt:

p
p n b a
n
e

( ) + + 1 2
(20)
where
n = number of bolt rows
The equation can be derived by dividing the total equiva-
lent length of the bolt group, based on the yield line pattern
of Mann and Morris (1979), by the total number of bolts in
the joint. The equivalent length is then used with the pry-
ing action procedure in the Manual. This equation accounts
for the prying effect on the bolts; however, the fact that the
outermost bolts take significantly more of the load than the
inner bolts is neglected.
EXPERIMENTAL RESEARCH
Ten independent research projects were located with ex-
perimental results on bolted tension connections. From
Fig. 6. Yield line pattern from Mann and Morris (1979).
Fig. 7. Circular yield line pattern from Zoetemeijer (1981).
Fig. 8. General load versus deformation curve.
093-116_ej2q_2011_2010-03r.indd 97 7/21/11 3:42 PM
Table 1. Geometry of Experimental Specimens
Specimen Geometry References
Type 1
Garrett (1977)
Grogan and Surtees (1999)
Hendrick and Murray (1983)
Moore and Sims (1986)
Packer and Morris (1977)
Pynnonen and Granstrom (1986)
Tawaga and Gurel (2005)
Zoetemeijer (1974)
Type 2
Garrett (1977)
Zoetemeijer (1981)
Type 3
Garrett (1977)
Type 4
Ghassemieh et al. (1983)
093-116_ej2q_2011_2010-03r.indd 98 7/21/11 3:42 PM
4. The ultimate strength. This is a point of interest for
structural integrity and ultimate strength calculations.
The deformation at this point is
u
and the load is P
u
.
Table A2 in Appendix A contains all of the loads described
for each specimen, where they were reported in the refer-
enced document. Many of the listed values are approximate
because they were read from graphs of the test data. The
experimental failure modes are also listed in Table A2. Most
of the specimens with two failure modes listed had the ulti-
mate strength limited by bolt rupture, but only after a large
deformation due to flange bending.
For the specimens with thick flanges, the bolts failed be-
fore the nonlinear part of the load-deformation curve was
reached. For these specimens, the bolt elongation contrib-
uted significantly to the total deformation.
For the specimens with thin flanges, the deformation at
ultimate strength was as much as 2 in. Under large deforma-
tions, the load-transfer mechanism changes from bending to
tension, which results in a tension load with a component
perpendicular to the axis of the bolt. This component is re-
sisted by the bolts in shear. Many of these tests resulted in
bolt fracture due to the applied tension combined with shear,
which was caused by large-deformation membrane action of
the fitting.
DEVELOPMENT OF PROPOSED
DESIGN METHOD
The purpose of this paper is to formulate a simple, accurate
and versatile method to design bolted flange connections. To
do this, the theoretical and experimental information pre-
sented by previous researchers will be analyzed.
The yield line solutions of Zoetemeijer (1974), Dranger
(1977), and Mann and Morris (1979) provide accurate results
for thin fittings where the limit state of bolt rupture is not
applicable. However, where thick flanges dictate that bolt
rupture is the controlling limit state, the yield line solutions
do not provide a method to calculate the prying force on
thebolt.
The method proposed by Thornton and Kane (1999) and
Muir and Thornton (2006) explicitly accounts for the prying
forces on the bolts; however, an equal amount of axial load is
assigned to each bolt. In reality, the outermost bolts will be
more highly stressed than the inner bolts, which could lead
to an unzipping action.
In this paper, a more refined solution has been developed,
where the forces are distributed according to the equivalent
length tributary to each bolt and the strength of each bolt is
evaluated independently. The equivalent tributary length is
calculated using existing yield line solutions.
The Component Method
Many different bolted flange configurations can be analyzed
by the yield line method; however, it would be cumbersome
for engineers to deal with a separate yield line pattern for
each different configuration. To simplify the design process,
the component method can be used, where single-bolt (lo-
cal) yield line patterns are assembled into a larger (global)
pattern for the entire bolt group. To do this, the engineer
simply selects a local pattern that is identical to each part
of the global pattern. The strength of each local pattern is
calculated and summed to get the total strength of the global
pattern.
In many cases, the local pattern will not be symmetrical
about the center of the bolt, and half-patterns can be used.
The strength of a half-pattern is simply half of the strength
of the whole pattern.
Stiffened Connections
If a flange is not adequate to carry the applied load, stiffen-
ers can be used to reinforce the joint as shown in Figure 9.
For stiffeners to be effective, they must be close enough to
the bolt to alter the yield line pattern. Using Drangers (1977)
yield line pattern, the stiffeners are effective if
x
s
< x (21)
where
x
s
= distance from the center of the bolt to the edge of
the stiffener
x = bc
Then, the strength can be determined by substituting x
s
for
x in Equation 15:

P F t
x
b
c
x
n y
s
s
+
j
(
,
\
,
(
2
(22)
Fig. 9. Yield line pattern for a stiffened flange in bending.
093-116_ej2q_2011_2010-03r.indd 99 7/21/11 3:42 PM
Equivalent Tributary Length Concept
The nominal strength from a given yield line pattern will
be equal to that of a straight yield line of length, p. Using
Equation 9 with d' = 0 and b' = b, the nominal strength for a
straight yield line is

P
F t p
b
n
y
2
2
(23)
To determine the equivalent tributary length of fitting, the
nominal strength of a given yield line solution will be set
equal to Equation 23 and solved for p. For the Dranger
(1977) pattern in Figure 5, the equivalent length is

p bc
d
4
(24)
For single-bolt connections, the equivalent tributary length
for the yield line solution of Mann and Morris (1979), shown
in Figure 6, is

p b a m + 2
(25)
The equivalent tributary length For Zoetemeijers (1981) cir-
cular pattern in Figure 7 is

p b
c
2
(26)
For the stiffened pattern in Figure 9, the equivalent tributary
length is

p x
cb
x
s s
s
+
j
(
,
\
,
(
2
(27)
For single-bolt connections, the yield line solution of
Zoetemeijer (1974), given by Equation 14, reduces to

p b a
z
+ 4 1 25 .
(28)
Selection of Proper Yield Line Solution
Because the yield line method is an upper-bound approach,
the pattern that gives the lowest load will provide results
closest to the true failure load. The normalized equivalent
lengths, p
e
/c, from the yield line solutions of Zoetemeijer
(1974, 1981), Dranger (1977), and Mann and Morris (1979)
are plotted against b/a in Figure 10. It can be seen that the
Mann and Morris (1979) solution results in the minimum
equivalent length for connections with high values of b/a,
and the Zoetemeijer (1981) solution produces the minimum
equival ent length only for connections with very low values
of b/a.
To simplify the design process, it is advantageous to use
only one of the available yield line patterns. Analysis of the
experimental deformations indicate that the yield line pat-
tern developed by Zoetemeijer (1974), shown in Figure 4,
is closest to the actual failure pattern. However, the skewed
yield lines are awkward to deal with if stiffeners are present,
and for most practical b/a ratios, the difference in strength of
the various yield line patterns is small.
The circular yield line pattern presented by Zoetemeijer
(1981) will control the design of fittings with large edge
distances, a. However, if a limit is placed on the b/a ratio,
this yield line pattern will never control the design. The
Zoetemeijer solution is equal to the Dranger (1977) solution
at b/a = 0.68; therefore, if a is limited to 1.47b for design
purposes, Zoetemeijers solution will never control. As a
slightly conservative (about 5%) limit, the prying action de-
sign procedure in the Manual (AISC, 2005a) can be used,
which limits a to a maximum of 1.25b.
When comparing the Dranger (1977) pattern to the Mann
and Morris (1979) pattern for stiffened flanges, the Dranger
pattern more accurately predicts the increase in strength
based on the distance from the bolt to the stiffener. This can
be verified by reviewing the projects that tested specimens
that were identical except for the addition of a stiffener: Pack-
er and Morriss (1977) specimens T6, T7 and T8; Moore and
Simss (1986) specimen T7. For these four specimens, the
Mann and Morris model predicted no increase in strength
due to the stiffeners; however, the average experimental load
increased by 32% compared to identical specimens with no
stiffeners. The Dranger model predicted a 37% increase due
to the stiffeners.
Due to the simplicity and the more accurate prediction of
the strength when stiffeners are present, the Dranger (1977)
yield line pattern is proposed here. A plot of p
min
/p
d
versus
b/a is shown in Figure 11, where p
d
is the tributary length
Fig. 10. Comparison of different yield line patterns.
093-116_ej2q_2011_2010-03r.indd 100 7/21/11 3:42 PM
for the Dranger yield line pattern and p
min
is the minimum
tributary length for the yield line patterns of Zoetemeijer
(1974), Dranger (1977), and Mann and Morris (1979). It is
seen that the Dranger solution is unconservative. However,
for most practical b/a ratios, the difference can be neglected
because the beneficial effects of strain hardening and mem-
brane action are not accounted for. Figure 11 also shows the
curve-fit equation, which can be used as a reduction factor
in design if the engineer wants to explicitly account for the
difference among the three different solutions. The curve fit
for the reduction factor is
C
r
= 1.0 0.11(b/a) + 0.019(b/a)
2
(29)
The coefficient of determination, R
2
, is 0.99, indicating a
very good fit. If the equivalent tributary length has been cal-
culated using Drangers solution, the minimum of the three
solutions can be approximated as
p
min
= p
d
C
r
(30)
where
p
min
= approximate minimum equivalent tributary
length per bolt
p
d
= equivalent tributary length per bolt calculated
according to the yield line pattern, developed
by Dranger (1977)
Joints with Bolt Rupture as the Controlling Limit State
In joints where the equivalent tributary length at one bolt is
larger than the remaining bolts in the joint, the bolt forces
will not be distributed equally. When bolt rupture is the con-
trolling limit state, the design procedure must account for
this. The component method accounts for the nonequal dis-
tribution of bolt forces by assigning the loads in proportion
to the tributary length at each bolt.
When bolt rupture controls the design, an additional com-
plication arises because deformation compatibility must be
maintained for all bolts in the joint. Under normal condi-
tions, when one of the bolts within the joint ruptures, it is
unlikely that the full yield line pattern has formed due to the
limited deformation. The local yield line pattern tributary to
the adjacent bolt will also be limited to the deformation at
bolt rupture. Because the internal energy at the yield lines is
proportional to the displacement, deformation compatibility
of the adjacent yield lines can be upheld by reducing the
strength in proportion to the deformation ratio,
r
/.

r a
T
T
1
(31)
where
r
= deformation at bolt rupture
= deformation at full yield line strength assuming
infinitely strong bolts
T
a
= strength of the fitting at bolt rupture (calculated
using Equation 1)
T
=1
= strength of the fitting assuming a full yield line
pattern forms without bolt rupture (calculated
using Equation 1 with = 1)
Using
r
as the deformation limit for the entire joint, the
total strength of the joint as controlled by the critical bolt is
P T
T
T
P
P
n cr
cr
cr
ei
e
+
j
(
,
,
\
,
(
(

j
(
,
,
\
,
(
(
,
,
,
]
]
]
]
1 1
1 ( ) ( )
max
(32)
where
T
cr
= strength of the fitting at the critical bolt (the
bolt with the largest equivalent tributary
length within the joint) (calculated using
Equation 1)
T
cr(=1)
= strength of the fitting at the critical bolt as-
suming a full yield line pattern forms without
bolt rupture (calculated using Equation 1 with
= 1)
P
ei
= summation of the equivalent tributary lengths
for all local yield line patterns within the joint
P
e(max)
= largest equivalent tributary length for all bolts
within the joint
Equation 32 provides a convenient way to deal with the de-
formation compatibility of the joint; however, when com-
pared to the test results of Ghassemieh et al. (1983), the
calculated strengths are very conservative. The conserva-
tism is due to the fact that the equation only accounts for the
flexural deformation of the fitting and neglects other defor-
mations within the joint, such as bolt elongation and shear
deformation of the fitting. As discussed in the section on
experimental research, bolt elongation can be a large portion
of the total joint deformation.
Fig. 11. Comparison of Dranger (1977) yield line
solution to the minimum of the Zoetemeijer (1974)
and Mann and Morris (1979) solutions.
093-116_ej2q_2011_2010-03r.indd 101 7/21/11 3:42 PM
group. If the yield line pattern is not symmetrical, two
half-patterns should be selected.
2. Calculate the strength of each bolt and fitting using
the prying action procedure in the Manual (AISC,
2005a), replacing p with p
e
; F
y
should be used in lieu
of F
u
unless large deformations are acceptable.
3. Repeat for all bolts in the bolt group.
4. Sum the individual strengths to get the total strength
of the bolt group.
The equivalent tributary length for the yield line pattern in
Figure 5 is

p bc
e
4
(33)
Where stiffeners are present, the equivalent length for the
pattern in Figure 9 is
p x if
cb
x
e s
s
+
j
(
,
\
,
(
2 x bc
s
< (34a)

p bc
e
4 if x bc
s

(34b)
Straight yield lines will be part of the yield line pattern when
the bolt spacing is less than 4 bc or the edge distance is less
than 2 bc. The equivalent tributary length per bolt is half
the distance between two bolts, p/2, or the distance from the
bolt to the end of the member, l
e
.
In the calculations for the equivalent tributary length, the
limit a 1.25b should be used. For connections subjected to
combined tension and shear, the bolt tension strength should
be reduced to account for the presence of shear.
As shown in the next section, serviceability design of
Due to the conservatism associated with Equation 32, it is
proposed that the strength of each bolt be evaluated indepen-
dently. Then, the total strength of the joint can be calculated
by summing the local capacities for the entire bolt group.
To account for the prying force on the bolt, the equivalent
tributary length, p
e
, is used in the prying action procedure
in the Manual in lieu of the tributary length, p. This proce-
dure provides nominal strengths that compare well with the
experimental loads, as discussed in the Experimental Vali-
dation section.
Large Bolt Spacings
If the distance between bolts, p, is greater than the equiv-
alent tributary length from Equation 24, two independent
yield lines will form for each bolt as shown in Figure 12a.
Figure 12b shows the same bolt pattern with a small bolt
spacing, where half-patterns form at each end and a straight
pattern forms between the bolts. Figure 13 shows a plot of
the equivalent length per bolt versus spacing between bolts.
The transition point between the two yield line patterns is at
a bolt spacing of 4 bc.
A similar problem occurs when a bolt is near the end of a
member. If the edge distance from the bolt to the end of the
member, l
e
, is less than 2 bc, a straight yield line will form
between the bolt and the end of the member.
PROPOSED DESIGN METHOD
The proposed design method consists of the following steps:
1. Select a valid yield line pattern local to each bolt in the

(a) (b)
Fig. 12. Effect of bolt spacing on the yield line
pattern: (a) large bolt spacing; (b) small bolt spacing. Fig. 13. Equivalent length per bolt versus spacing between bolts.
093-116_ej2q_2011_2010-03r.indd 102 7/21/11 3:42 PM
connections that can allow only very small deformations
should be based on 60% of the nominal load calculated us-
ing F
y
with the proposed design method. However, for most
standard connections, a reduction for stiffness is not re-
quired because a 4-in. deformation allowance is not uncom-
mon in determining the nominal strength of connections for
various limit states. For example, in the AISC Specification
for Structural Steel Buildings (AISC, 2005b), the nominal
strength for bearing strength at bolt holes is based on a de-
formation limit of 4 in., with an increase in the nominal
strength available if more deformation is allowed.
EXPERIMENTAL VALIDATION
The proposed design method was compared to the results of
59 tests from 10 independent research projects. The experi-
mental results are shown in Table A2 of Appendix A. Table
A3 in Appendix A shows the calculated nominal strengths
and predicted failure modes for all of the specimens. Table
A3 also shows the test-to-calculated ratios for each available
data point on the experimental curves.
The procedure outlined in the proposed design method
was used to calculate the nominal strength of each specimen
using the actual yield strengths and the ultimate strengths
reported in the referenced documents. Several of the refer-
enced documents reported the yield strength of the tested
material, but omitted the ultimate strength; therefore, there
were fewer experimental data points to compare with the
ultimate strength calculations.
For each specimen, the nominal strength at each local
yield line pattern was calculated using Equation 35, with the
nominal value of t
c
calculated without the resistance factor,
as expressed in Equations 36a and 36b for the yield and ulti-
mate strength solutions, respectively.

T r
t
t
r
e t
c
t
j
(
,
\
,
(
+
( )
2
1
(35)
t
r b
p F
c
t
e y
4
(36a)
t
r b
p F
c
t
e u
4
(36b)
where
r
t
= strength of tested bolt in tension
Then, the individual strengths of all local yield line patterns
within the joint were summed to get the total strength of the
joint.
The statistical results are summarized in Table 2, which
provides the number of specimens with adequate data to be
included in the results, the average, the standard deviation
and the low values for the 95% and 99% confidence inter-
vals. Note that P
p
, P
s
, P
4
and P
u
are defined in the section
on Experimental Research, and P
ny
and P
nu
are the nomi-
nal loads calculated with the yield strength and ultimate
strength of the fitting, respectively.
The results show that the load at 4-in. deformation, P
4
,
can be accurately predicted using F
y
with the proposed de-
sign method. From column 4 in Table 2, the average test-to-
predicted ratio for the 23 specimens is 1.12, and the standard
deviation is 0.262. The low values for the 95% and 99% con-
fidence intervals are 1.01 and 0.976, respectively.
The ultimate loads can be accurately predicted using F
u

with the proposed design method. However, the deforma-
tions at ultimate strength can be very largeTable A2 in
Appendix 2 shows experimental deformations greater than
1 in. for several specimens at the maximum test load. From
Table 2. Summary of Calculation Results
[1]
Using F
y
Using F
u
P
P
p
ny
[2]
P
P
s
ny
[3]
P
P
ny
4
[4]
P
P
u
ny
[5]
P
P
p
nu
[6]
P
P
s
nu
[7]
P
P
nu
4
[8]
P
P
u
nu
[9]
Number of specimens 43 30 23 52 14 11 8 12
Average 0.675 0.932 1.12 1.61 0.490 0.685 0.763 1.13
Standard deviation 0.173 0.184 0.262 0.378 0.157 0.148 0.170 0.269
95% confidence interval
(low value)
0.623 0.866 1.01 1.51 0.407 0.597 0.646 0.976
99% confidence interval
(low value)
0.607 0.845 0.976 1.48 0.381 0.570 0.609 0.929
093-116_ej2q_2011_2010-03r.indd 103 7/21/11 3:42 PM
Example 2
Determine the equivalent tributary length for each bolt in
Figure 15.
For bolt 1: Assume l bc
e
< 2 . The equivalent length is the
distance from bolt 1 to the end of the member plus half of the
distance between bolts 1 and 2.
p l
p
e e
+
2
For bolt 2: The equivalent length is half of the equivalent
length from Equation 33 plus half of the distance between
bolts 1 and 2.
p bc
p
e
+ 2
2
Example 3
Figure 16.
For bolt 1: Assume l bc
e
< 2 and x
s
< x. The equivalent
length is the distance from bolt 1 to the end of the member
plus half of the equivalent length from Equation 34a.
p l x
bc
x
e e s
s
= + +
1
1
For bolt 2: Assume x
s
< x. The equivalent length is half of
the equivalent length from Equation 34a plus half of the dis-
tance between bolts 2 and 3.
p
p
x
bc
x
e s
s
= + +
23
2
2
2
column 9 in Table 2, the average test-to-predicted ratio for
the 12 specimens is 1.13, and the standard deviation is 0.269.
The low values for the 95% and 99% confidence intervals
are 0.976 and 0.929, respectively.
A comparison of columns 4 and 2 of Table 2 indicates that
the load at the proportional limit is about 60% of the load
at -in. deformation. Based on this, serviceability design
of connections that can allow only very small deformations
may be based on 60% of the nominal load calculated using
F
y
with the proposed design method.
EXAMPLES
Example 1
Figure 14.
bolts 1 and 2.
p bc
p
e
= + 2
2
12
For bolt 2: The equivalent length is half of the distance be-
tween bolts 1 and 2 plus half of the distance between bolts
2 and 3.
p
p p
e
= +
12 23
2 2
bolts 2 and 3.
p bc
p
e
= + 2
2
23
Fig. 14. Example 1. Fig. 15. Example 2.
093-116_ej2q_2011_2010-03r.indd 104 7/21/11 3:42 PM
Fig. 17. Example 4.
Use a = 1.36 in.
=
=
b 2 56
0 75
2
2 18
.
.
.
in.
in.
in.
=
=
a 1 36
0 75
2
1 74
.
.
.
in.
in.
in.
+
=

=
=
b
a
2 18
1 74
1 25
. .
. .
.
in
in
c = b + a = 2.56 in. + 1.36 in. = 3.92 in.
For bolts at row 1,
P
eff
= 3.00 in.
t
Bb
p F
c
eff y
=

=
( )( )( )
( )( )
=
4 44
4 44 29 8 2 18
3 50
1
.
. . . .
.
kips in
in. ksi
339 in.
=

=
=
1
1
0 8125
3
0 729
d
p
eff
.
.
in.
in.
bolts 2 and 3.
p bc
p
e
= + 2
2
23
Example 4
Figure 17.
For bolts 1 and 2: Assume x
s
< x. The equivalent length is
half of the equivalent length from Equation 33 plus half of
the equivalent length from Equation 34a.
p bc x
bc
x
e s
s
= + + 2
Example 5
Determine the available LRFD strength of the connection in
Figure 18 for the limit states of bolt rupture and beam flange
bending. The beam is a W2155 of A992 material. Bolts are
w-in.-diameter A325 with m-in.-diameter holes. The beam
gage, g, is 52 in.
B = r
n
= 29.8 kips
t
f
= 0.522 in.
t
w
= 0.375 in.
b
f
= 8.22 in.
b =

=
5 5 0 375
2
2 56
. .
.
in. in.
in.
a =

=
8 22 5 5
2
1 36
. .
.
in. in.
in.
For design purposes, a must not be greater than 1.25b.
a < 1.25b
1.36 in. < (1.25)(2.56 in.)
Fig. 16. Example 3. Fig. 18. Example 5hanger connection without stiffeners.
093-116_EJ2Q_2011_2010-03R.indd 105 8/11/11 10:16 AM
Use = 1.00.
T B
t
t
n
c
j
(
,
\
,
(
+
( )
( )
j
(
,
\
,
(
+
2
2
1
29 8
0 522
1 10
1 0 .
. .
. .
kips
in
in
.. .
.
829 1 00
12 3
( )( ) ,
]
]
kips
For bolts at row 3,
p x
p
eff
+
( )( ) +
2
2
2 3 17
6 5
2
9 59
. .
. .
. .
in
in
in
t
Bb
p F
c
eff y
( )( )( )
( )( )
4 44
4 44 29 8 2 18
9 59 50
.
. . . .
.
kips in
in. ksi
0 776 . in.

1
1
0 8125
9 59
0 915
d
p
eff
.
.
.
in.
in.

+ ( )
j
(
,
\
,
(

,
,
,
]
]
]
]
( ) + ( )

1
1
1
1
0 915 1 1 25
0 776
0 5
2
t
t
c
. .
. .
.
in
222
1
0 588
2
in.
.
j
(
,
\
,
(

,
,
,
]
]
]
]
T B
t
t
n
c
j
(
,
\
,
(
+
( )
( )
j
(
,
\
,
(
+
2
2
1
29 8
0 522
0 776
1 .
. .
. .
kips
in
in
00 915 0 588
20 7
. .
.
( )( ) ,
]
]
kips
The available load for the serviceabili ty limit state is
P
n
= (2)(7.27 kips) + (4)(12.3 kips + 20.7 kips) = 146 kips
If the ultimate strength, F
u
= 65 ksi, is used in the design
procedure, the available load for the strength limit state is
P
n

+ ( )
j
(
,
\
,
(

,
,
,
]
]
]
]
( ) + ( )

1
1
2
1
1
0 729 1 1 25
1 39
0 52
t
t
c
. .
. .
.
in
22
2
1
3 71 1 00
in.
. .
j
(
,
\
,
(

,
,
,
]
]
]
]
>
Use = 1.00.
T B
t
t
n
c
j
(
,
\
,
(
+
( )
( )
j
(
,
\
,
(
+
2
1
29 8
0 522
1 39
2
1 0 .
. .
. .
kips
in
in
.. .
.
729 1 00
7 27
( )( ) ,
]
]
kips
For bolts at row 2,
p
eff
+
3
2
6 5
2
4 75
in. in
in
. .
. .
t
Bb
p F
c
eff y
( )( )( )
( )( )
4 44
4 44 29 8 2 18
4 75 50
.
. . . .
.
kips in
in. ksi
1 10 in. .

1
1
0 8125
4 75
0 829
d
p
eff
.
.
.
in.
in.

+ ( )
j
(
,
\
,
(

,
,
,
]
]
]
]
( ) + ( )

1
1
1
1
0 829 1 1 25
1 10
0 52
2
t
t
c
. .
. .
.
in
22
1
1 84
2
in.
.
j
(
,
\
,
(

,
,
,
]
]
]
]
093-116_ej2q_2011_2010-03r.indd 106 7/21/11 3:42 PM

Example 6
If more strength is required for the connection in Example5,
stiffeners can be added as shown in Figure 19. Determine
the strength of the bolts and beam flange.
For bolts at row 1,
P
eff
= 3.00 in.
T
n
7 27 . kips (from Example 5)
For bolts at row 2,
x
s
= 2.50 in.
x bc
( )( )
2 56 3 92
3 17
. . .
.
in. in
in.
Thus, x
s
< x; therefore,
p
p
x
bc
x
eff s
s
+ +
+ +
( )( )
2
3
2
2 50
2 56 3 92
2 50
8
in.
in.
in in
in
.
. . . .
. .
.001 in.
t
Bb
p F
c
eff y
( )( )( )
( )( )
4 44
4 44 29 8 2 18
8 01 50
.
. . . .
.
kips in
in. ksi
0 849 . in.

1
1
0 8125
8 01
0 899
d
p
eff
.
.
.

in

+ ( )
j
(
,
\
,
(

,
,
,
]
]
]
]
( ) + ( )

1
1
1
1
0 899 1 1 25
0 849
0 5
2
t
t
c
. .
. .
.
in
222
1
0 813
2
in.
.
j
(
,
\
,
(

,
,
,
]
]
]
]
Fig. 19. Example 6hanger connection with stiffeners.

T B
t
t
n
c
j
(
,
\
,
(
+
( )
( )
j
(
,
\
,
(
+
2
2
1
29 8
0 522
0 849
1 .
. .
. .
kips
in
in
00 899 0 813
19 5
. .
.
( )( ) ,
]
]
kips
For bolts at row 3,
x
s
= 3.50 in.
x
s
= 3.17 in.
Thus, x
s
> x; therefore,
p x
eff

( )( )
4
4 3 17
12 7
. .
. .
in
in
t
Bb
p F
c
eff y
( )( )( )
( )( )
4 44
4 44 29 8 2 18
12 7 50
.
. . . .
.
kips in
in. ksi
0 674 . in.
093-116_ej2q_2011_2010-03r.indd 107 7/21/11 3:42 PM

1
1
0 8125
12 7
0 936
d
p
eff
.
.
.
in.
in.

+ ( )
j
(
,
\
,
(

,
,
,
]
]
]
]
( ) + ( )

1
1
1
1
0 936 1 1 25
0 674
0 5
2
t
t
c
. .
. .
.
in
222
1
0 317
2
in.
.
j
(
,
\
,
(

,
,
,
]
]
]
]
T B
t
t
n
c
=
j
(
,
\
,
(
+
( )
= ( )
j
(
,
\
,
(
+
2
2
1
29 8
0 522
0 674
1 .
. .
. .
kips
in
in
00 936 0 317
23 2
. .
.
( )( ) ,
]
]
= kips
The available load for the serviceability limit state is
P
n
If the ultimate strength, F
u
= 65 ksi, is used in the design
procedure, the available load for the strength limit state is
P
n
CONCLUSIONS
A method has been proposed to calculate the tension
strength of bolted flange connections, which includes the
effects of prying action. The proposed design procedure,
based on yield line theory, is simple, accurate and versatile.
It can be used to calculate the strength of many different
connection configurations, including stiffened connections,
connections with large bolt spacing and connections close to
the end of the member.
The bolt forces within a group are distributed according to
the equivalent length tributary to each bolt, and the strength
of each bolt is evaluated independently. The total strength of
the joint is then calculated by summing the nominal strength
at each bolt for the entire bolt group.
The calculated strengths were compared to the results of
59 tests from 10 independent research projects, and the pro-
posed design method, which uses the yield strength of the
fitting, was shown to be accurate for a deformation limit of
approximately 4 in. The ultimate strength of the fitting can
be determined by using the proposed design procedure with
the ultimate strength, F
u
, of the fitting. However, the defor-
mations at the ultimate strength can be large.
SYMBOLS
B Availa ble tension per bolt
C
r
Curve fit for the reduction factor is
F
y
Specified minimum yield strength of the fitting
F
u
Specified minimum tensile strength of the fitting
L
i
Length of yield line i
M
pi
Plastic moment capacity of yield line i
P Applied load
P
ny
Nominal load calculated with the yield strength of
the fitting
P
nu
Nominal load calculated with the ultimate strength
of the fitting
P
p
Experimental load at the proportional limit
P
s
Experimental load at the nonlinear transition
point on the load-deformation curve
P
u
Experimental load at ultimate failure
P

Experimental load at -in. deformation
T
a
Available tensile strength of fitting
T
cr
Strength of the fitting at the critical bolt (the bolt
with the largest equivalent tributary length within
the joint) (calculated using Equation1)
T
cr (=1)
Strength of the fitting at the critical bolt
assuming a full yield line pattern forms without
bolt rupture (calculated using Equation 1 with
=1)
T
=1
Strength of the fitting assuming a full yield line
pattern forms without bolt rupture (calculated
using Equation 1 with =1)
a Distance from the bolt centerline to the edge of
the fitting, but 1.25b for calculations using the
proposed design method
b Distance from bolt centerline to the face of the
web
d
b
Bolt diameter
d Width of the hole along the length of the fitting
l
e
Edge distance from the bolt to the end of the
member
m
p
Plastic moment capacity per unit length of the
fitting
n Number of bolt rows
p Spacing between bolts
093-116_ej2q_2011_2010-03r.indd 108 7/21/11 3:42 PM
p
z
Equivalent tributary length per bolt calculated
according to the yield line pattern developed by
Zoetemeijer (1974)
r
t
Strength of tested bolt in tension
r
n
Nominal strength of bolt in tension
t Thickness of the fitting, in.
x bc
x
s
Distance from the center of the bolt to the edge of
the stiffener
Virtual displacement
i
Virtual rotation of yield line i
p
Experimental deformation at the proportional
limit
s
Experimental deformation at the nonlinear
transition point on the load-deformation curve
u
Experimental deformation at ultimate failure
p
c
according to the circular yield line pattern
developed by Zoetemeijer (1981)
p
d
Dranger (1977)
p
e
Equivalent length of fitting tributary to the bolt
in question for connection type 4, the equivalent
length of fitting tributary to the bolt farthest from
the end
p
el
For connection type 4, the equivalent length of
fitting tributary to the bolt closest to the end
p
m
Mann and Morris (1979)
p
min
Minimum equivalent tributary length per bolt;
minimum of p
d,
p
z
and p
m
p
min
Approximate minimum equivalent tributary
length per bolt calculated with Equation 30
p
s
according to the stiffened yield line pattern in
Figure 9
093-116_ej2q_2011_2010-03r.indd 109 7/21/11 3:42 PM
APPENDIX A
Table A1. Specimen Properties
Specimen
F
y
(ksi)
F
u
(ksi)
t
(in.)
b
(in.)
a
(in.)
p
e
(in.)
p
el
(in.)
d
b
(in.)
d
(in.)
r
t
(kips)
Notes
Garrett (1977)
1 41.2 66.5 0.465 1.62 1.51 7.00 0.875 0.938 54.1
2 41.2 66.5 0.465 1.62 1.51 9.00 0.875 0.938 54.1
3 41.2 66.5 0.465 1.62 1.51 9.00 0.875 0.938 54.1
TH-1 43.5 65.9 0.5 1.5 2.92 5.62 2.08 0.625 0.688 27.1
TH-2 38.7 70.1 0.5 2.5 2.92 8.23 2.08 0.625 0.688 27.1
TH-3 45.4 66.5 1 1.5 2.92 5.00 2.08 0.625 0.688 27.1
TH-4 43.1 73.9 1 2.5 2.92 10.3 2.08 0.625 0.688 27.1
TH-5 44.9 72.5 0.75 1.5 2.92 5.62 2.08 0.625 0.688 27.1
TH-6 37.7 65.2 0.75 2.5 2.92 8.23 2.08 0.625 0.688 27.1
E1 42.8 0.559 2.59 2.24 10.0 1.18 1.30 64.4
E15 42.2 0.559 2.59 2.24 14.4 1.18 1.30 64.4
1 38.3 0.778 2.52 3.26 9.95 1.38 1.44 133
2 34.6 0.813 2.50 4.54 10.7 1.38 1.44 133
3 39.7 0.718 2.54 3.25 9.57 1.13 1.19 89.5
4 39.7 0.718 2.54 3.25 10.14 1.50 1.56 159
T1 44.8 0.268 1.57 1.30 6.22 0.630 0.709 25.2
T8 44.8 0.268 1.57 1.30 8.54 0.630 0.709 25.2
T1 42.9 0.268 1.58 1.12 6.04 0.630 0.811 36.7
T2 43.0 0.268 1.58 1.12 6.04 0.630 0.811 36.7
T3 42.5 0.354 1.56 1.12 6.01 0.630 0.811 36.7
T4 44.7 0.268 1.73 0.96 7.26 0.630 0.811 36.7
T5 43.6 0.528 1.49 1.12 5.86 0.630 0.811 36.7
T6 43.2 0.268 1.58 1.12 8.30 0.630 0.811 36.7
T7 43.9 0.268 1.58 1.12 8.30 0.630 0.811 36.7
T8 44.8 0.268 1.58 1.12 8.30 0.630 0.811 36.7
1 46.1 68.4 0.366 1.38 1.65 6.07 0.630 0.709 31.1
5 46.1 68.4 0.366 1.74 1.30 6.57 0.630 0.709 31.1
11 46.1 68.4 0.366 1.74 1.30 6.57 0.945 1.02 71.6
15 44.8 64.4 0.551 2.19 3.54 9.07 0.945 1.02 71.6
21 44.8 64.4 0.551 3.42 2.32 10.8 0.945 1.02 71.6
093-116_ej2q_2011_2010-03r.indd 110 7/21/11 3:42 PM
Table A1. Specimen Properties (continued)
Specimen
F
y
(ksi)
F
u
(ksi)
t
(in.)
b
(in.)
a
(in.)
p
e
(in.)
p
el
(in.)
d
b
(in.)
d
(in.)
r
t
(kips)
Notes
T-N 41.8 61.9 0.394 1.44 1.38 6.03 0.787 0.866 50.9
Zoetemeijer (1981)
1 36.4 0.457 2.21 2.36 13.9 0.787 0.866 50.9
2 36.4 0.457 3.18 1.38 18.2 0.787 0.866 50.9
4 36.4 0.457 2.21 2.36 12.8 0.787 0.866 50.9
5 36.4 0.457 3.18 1.38 9.54 0.787 0.866 50.9
6 36.4 0.457 3.18 1.38 15.3 0.630 0.709 32.9
7 36.4 0.457 3.18 1.38 15.8 0.787 0.866 50.9
8 40.6 0.492 4.39 1.36 20.7 0.945 1.02 67.5
9 40.6 0.492 2.19 3.56 13.7 0.945 1.02 67.5
10 40.6 0.492 4.39 1.36 20.7 0.945 1.02 67.5
11 40.6 0.492 2.19 3.56 14.2 0.945 1.02 67.5 1
12 40.6 0.492 4.39 1.36 27.6 0.945 1.02 67.5
13 40.6 0.492 2.19 3.56 13.7 0.945 1.02 67.5
14 40.6 0.492 2.19 3.56 17.5 0.945 1.02 67.5 1
15 40.6 0.492 4.39 1.36 27.6 0.945 1.02 67.5
Zoetemeijer (1974)
5 37.7 0.315 1.68 0.965 5.78 0.787 0.866 39.4
6 37.7 0.315 1.28 1.13 5.09 0.787 0.866 39.4
7 38.8 0.335 1.77 1.26 6.21 0.787 0.866 39.4
8 41.8 0.492 1.77 1.22 6.17 0.787 0.866 41.1
9 41.8 0.492 1.38 1.13 5.29 0.787 0.866 41.1
10 39.2 0.906 1.61 1.26 5.88 0.787 0.866 38.9
11 43.5 0.591 1.87 1.12 6.29 0.787 0.866 41.1
12 43.5 0.669 1.69 1.26 6.03 0.787 0.866 38.3
13 37.7 0.315 1.68 0.97 5.78 0.787 0.866 39.4
14 37.7 0.315 1.68 0.97 5.78 0.787 0.866 39.4
20 30.5 0.571 2.48 1.12 7.54 0.787 0.866 40.7
21 30.5 0.571 2.48 1.12 7.54 0.787 0.866 40.7
22 30.5 0.571 2.48 1.12 7.54 0.787 0.866 37.2
23 30.5 0.571 2.48 1.12 7.54 0.787 0.866 37.2
1. Four bolts per bolt row.
093-116_ej2q_2011_2010-03r.indd 111 7/21/11 3:42 PM
Table A2. Experimental Results
Specimen
P
p
(kips)
P
s
(kips)
P
4
(kips)
P
u
(kips)
p
(in.)
n
(in.)
u
(in.)
Failure
Mode
Notes
Garrett (1977)
1 56 82 92 164 0.08 0.18 1.39 F
2 72 102 106 174 0.12 0.23 0.74 F
3 85 97 117 239 0.13 0.17 1.29 F
TH-1 80 105 0.002 0.006 B
TH-2 40 110 140 0.001 0.016 > 0.1 F B
TH-3 165 200 0.006 N 1
TH-4 150 175 195 0.007 0.014 0.03 B
TH-5 130 150 170 0.006 0.012 > 0.02 F B
TH-6 70 140 0.001 0.011 N 2
E1 90 124 248 F
E15 112 169 292 F
1 110 160 200 0.003 0.021 0.065 N 1
2 120 160 200 0.012 0.033 0.088 N 1
3 60 200 0.023 0.22 N 1
4 110 170 200 0.012 0.056 0.14 N 1
T1 22 27 31 60 0.10 0.18 > 0.7 O
T8 34 40 43 74 0.16 0.22 > 0.6 O
T1 22.5 63.0 F B
T2 22.1 62.8 F B
T3 36.0 69.3 F B
T4 27.0 45.5 F O
T5 63.0 103 O B
T6 31.5 73.6 F O
T7 31.5 73.6 F O
T8 29.3 67.7 F O
1 51 68 83 97.7 0.03 0.08 > 0.6 F B
5 22 50 61 90.0 0.05 0.08 > 0.6 F B
11 30 57 94 151 < 0.01 0.05 > 0.9 F
15 72 110 130 212 < 0.01 0.03 > 0.9 F B
21 191 F B
093-116_ej2q_2011_2010-03r.indd 112 7/21/11 3:42 PM
Table A2. Experimental Results (continued)
Specimen
P
p
(kips)
P
s
(kips)
P
4
(kips)
P
u
(kips)
p
(in.)
n
(in.)
u
(in.)
Failure
Mode
Notes
T-N 61 74 94 97 0.024 0.063 0.32 N 3
Zoetemeijer (1981)
1 167 F
2 56 76 79 144 0.08 0.24 1.2 F
4 161 F
5 135 F
6 117 B
7 133 F
8 50 80 79 183 0.09 0.18 2.4 B
9 90 150 140 244 0.04 0.35 2.0 B
10 40 65 68 183 0.08 0.20 2.2 O
11 97 150 140 266 0.06 0.35 1.4 B
12 150 O
13 130 190 221 0.08 B
14 250 290 280 300 0.12 0.43 0.59 O
15 159 O
Zoetemeijer (1974)
5 22 32 45 49.5 0.004 0.02 > 0.2 F 4
6 32 45 63 67.4 0.01 0.03 > 0.2 F 4
7 26 40 48 67.4 0.01 0.03 > 0.2 F 4
8 135 F B
9 135 F B
10 148 F B
11 126 F B
12 153 F B
13 22 29 36 49.5 0.02 0.04 > 0.2 F 4
14 25 30 36 40.5 0.04 0.05 > 0.2 F 4
20 120 F B
21 43 63 84 103 0.02 0.04 > 0.2 F B 4
22 111 F B
23 128 F B
Notes
1. Maximum test load was 200 kips, which was the machine capacity.
2. Test result for ultimate load was not available; 140-kip load was taken
from finite element model.
3. Test was stopped at a load of 97 kips.
4. Loads at 4-in. deformation were conservatively read from the highest
graphed deformations, which were between 0.12 and 0.16 in.
Failure modes
N: No failure
F: Flange bending
B: Bolt rupture
O: Other
093-116_ej2q_2011_2010-03r.indd 113 7/21/11 3:42 PM
Table A3. Calculation Results
Specimen
Nominal
Strength (kips)
Predicted
Failure Mode
P
P
p
ny
P
P
s
ny
P
P
ny
4
P
P
u
ny
P
P
p
nu
P
P
s
nu
P
P
nu
4
P
P
u
nu
Notes
Using
F
y
Using
F
u
Using
F
y
Using
F
u
Garrett (1977)
1 98.6 159 F F 0.57 0.83 0.93 1.66 0.35 0.52 0.58 1.03
2 129 176 F F B 0.56 0.79 0.82 1.35 0.41 0.58 0.60 0.99 2
3 129 176 F F B 0.66 0.75 0.91 1.85 0.48 0.55 0.66 1.36 2
TH-1 116 144 F B F B 0.69 0.91 0.56 0.73 1
TH-2 81.8 111 F B F B 0.49 1.34 1.71 0.36 0.99 1.26 1
TH-3 204 217 B B 0.81 0.76 1
TH-4 177 195 B B 0.85 0.99 1.10 0.77 0.90 1.00 1
TH-5 182 201 B B 0.71 0.82 0.93 0.65 0.75 0.85 1
TH-6 125 167 F B B 0.56 0.42 1
E1 126 F 0.71 0.98 1.97
E15 181 F 0.62 0.93 1.61
1 233 F 0.47 0.69
2 252 F 0.48 0.63
3 186 F 0.32
4 215 F 0.51 0.79
T1 30.0 F 0.73 0.90 1.03 2.00
T8 41.8 F 0.81 0.96 1.03 1.77 1
T1 27.4 F 0.82 2.30
T2 27.5 F 0.80 2.28
T3 47.9 F 0.75 1.45
T4 31.1 F 0.87 1.46
T5 108 F B 0.58 0.95
T6 38.7 F 0.81 1.90 1
T7 39.3 F 0.80 1.87 1
T8 40.1 F 0.73 1.69 1
1 66.1 98.1 F F 0.77 1.03 1.26 1.48 0.52 0.69 0.85 1.00
5 54.0 80.1 F F 0.41 0.93 1.13 1.67 0.27 0.62 0.76 1.12
11 59.1 87.8 F F 0.51 0.96 1.59 2.55 0.34 0.65 1.07 1.72
15 135 194 F F 0.53 0.81 0.96 1.57 0.37 0.57 0.67 1.09
21 95.4 137 F F 2.00 1.39
093-116_ej2q_2011_2010-03r.indd 114 7/21/11 3:42 PM
Table A3. Calculation Results (continued)
Specimen
Nominal
Strength (kips)
Predicted
Failure Mode
P
P
p
ny
P
P
s
ny
P
P
ny
4
P
P
u
ny
P
P
p
nu
P
P
s
nu
P
P
nu
4
P
P
u
nu
Notes
Using
F
y
Using
F
u
Using
F
y
Using
F
u
T-N 69.5 103 F F 0.88 1.06 1.35 0.59 0.72 0.91
Zoetemeijer (1981)
1 112 F 1.49 1, 3
2 96.8 F 0.58 0.79 0.82 1.49 1
4 103 F 1.56 1
5 49.5 F 2.73 1
6 74.2 B 1.58
7 83.5 F 1.59
8 101 F 0.50 0.79 0.78 1.81 1
9 152 F 0.59 0.99 0.92 1.61 3
10 101 F 0.40 0.64 0.67 1.81 1
11 157 F 0.62 0.96 0.89 1.69
12 133 B 1.13 1, 3
13 152 F 0.86 1.25 1.45 1, 3
14 195 F 1.28 1.49 1.44 1.54 1
15 133 B 1.20 1, 3
Zoetemeijer (1974)
5 31.2 F 0.71 1.03 1.44 1.59
6 39.2 F 0.82 1.15 1.61 1.72
7 36.5 F 0.71 1.10 1.32 1.85
8 84.3 F 1.60
9 99.8 F 1.35
10 155 B 0.95
11 115 F B 1.10
12 126 F B 1.21
13 31.2 F 0.71 0.93 1.15 1.59
14 31.2 F 0.80 0.96 1.15 1.30
20 67.9 F 1.77
21 67.9 F 0.63 0.93 1.24 1.52
22 67.9 F 1.63
23 67.9 F 1.89
Notes
1. x
s
< x
2. x
s
> x
3. Theory indicates circular yield line controls the design. This was accounted for in the listed values.
093-116_ej2q_2011_2010-03r.indd 115 7/21/11 3:42 PM
Moore, D.B. and Sims, P.A.C. (1986), Preliminary Inves-
tigations into the Behavior of Extended End-Plate Steel
Connections with Backing Plates, Journal of Construc-
tional Steel Research, pp. 95122.
Muir, L.S. and Thornton, W.A. (2006), Connections,
Ch.3, Structural Steel Designers Handbook, R.L. Brock-
enbrough and F.S. Merritt, Eds., McGraw-Hill, New York.
Packer, J.A. and Morris, L.J. (1977), A Limit State Method
for the Tension Region of Bolted Beam-Column Connec-
tions, The Structural Engineer, Vol. 5, No. 10.
Pynnonen, J. and Granstrom, A. (1986), Beam-to-Column
Connections with Backing Plates, Swedish Institute of
Steel Construction Report 86:6, April.
SCI (1995), Joints in Steel Construction-Moment Connec-
tions, The Steel Construction Institute, Ascot, Berkshire,
UK.
Tawaga, H. and Gurel, S. (2005), Strength Evaluation of
Bolted Moment Connections Stiffened with Channels,
Handout for poster session at the North American Steel
Construction Conference.
Thornton, W.A. (1985), Prying ActionA General Treat-
ment, Engineering Journal, AISC, Vol. 22, No. 2, pp.
6775.
Thornton, W.A. (1992), Strength and Serviceability of
Hanger Connections, Engineering Journal, Vol. 29, No.
4, pp. 145149.
Thornton, W.A. and Kane, T. (1999), Design of Connec-
tions for Axial, Moment and Shear Forces, Ch. 2, Hand-
book of Structural Steel Connection Design and Details,
A.R. Tamboli, Ed., McGraw-Hill, New York.
Young, W.C. (1989), Roarks Formulas for Stress and Strain,
6th ed., McGraw-Hill, p. 207.
Zoetemeijer, P. (1974), A Design Method for the Tension
Side of Statically Loaded, Bolted Beam-to-Column Con-
nections, Heron, Vol. 20, No. 1.
Zoetemeijer, P. (1981), Semi-Rigid Bolted Beam-to-Beam
Column Connections with Stiffened Column Flanges and
Flush End Plates, Joints in Structural Steelwork, Proc.
International Conference, Teesside Polytechnic, April
69, John Wiley and Sons.
REFERENCES
AISC (2005a), Steel Construction Manual, 13th ed., Ameri-
can Institute of Steel Construction, Chicago, IL.
AISC (2005b), Specification for Structural Steel Buildings,
American Institute of Steel Construction, Chicago, IL.
Astaneh, A. (1985), Procedure for Design and Analysis of
Hanger-Type Connections, Engineering Journal, AISC,
Vol. 22, No. 2, pp. 6366.
CEN (2005), Eurocode 3: Design of Steel Structures-
Part 18: Design of Joints, European Committee for
Standardization.
Douty, R.T. and McGuire W. (1965), High Strength Bolted
Moment Connections, Journal of the Structural Divi-
sion, ASCE, Vol. 91, No. ST2, pp. 101128.
Dranger, T.S. (1977), Yield Line Analysis of Bolted Hang-
ing Connections, Engineering Journal. AISC, Vol. 14,
No. 3.
Garrett, J.L. (1977), An Investigation of High Strength
Bolted Tension Connections, Masters Thesis, Iowa State
University.
Ghassemieh, M., Kukreti, A. and Murray, T.M. (1983), In-
elastic Finite Element Analysis of Stiffened End Plate
Moment Connections, Report FSEL/AISC 83-02, School
of Civil Engineering and Environmental Science, Univer-
sity of Oklahoma, June.
Grogan, W. and Surtees, J.O. (1999), Experimental Be-
havior of End Plate Connections Reinforced with Bolted
Backing Angles, Journal of Constructional Steel Re-
search, Vol. 50, pp. 7196.
Hendrick, A. and Murray, T.M. (1983), Column Web and
Flange Strength at End-Plate Connections, Report FSEL/
AISC 83-01, School of Civil Engineering and Environ-
mental Science, University of Oklahoma, February.
Hognestad, E. (1953), Yield Line Theory for the Ultimate
Flexural Strength of Reinforced Concrete Slabs, Journal
of the American Concrete Institute, Vol. 24, No. 7.
Johansen, K.W. (1962), Yield Line Theory, Cement and Con-
crete Association, London, England.
Kulak, G.L., Fisher, J.W. and Struik, J.H.A. (1987), Guide to
Design Criteria for Bolted and Rivited Joints, John Wiley
and Sons, New York.
Mann, A.P. and Morris, L.J. (1979), Limit Design of Ex-
tended End-Plate Connections, Journal of the Structural
Division, ASCE, March, pp. 511526.
093-116_ej2q_2011_2010-03r.indd 116 7/21/11 3:42 PM

A Yield Line Component Method For Bolted Flange Connections-1

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

A Yield Line Component Method For Bolted Flange Connections-1

Încărcat de

Drepturi de autor:

Formate disponibile

ENGINEERING JOURNAL / SECOND QUARTER / 2011 / 93

A Yield Line Component Method

093-116_ej2q_2011_2010-03r.indd 106 7/21/11 3:42 PM

Fig. 19. Example 6hanger connection with stiffeners.

S-ar putea să vă placă și